On Wed, Feb 26, 2020 at 01:26:09PM -0800, Kuppuswamy Sathyanarayanan wrote:
> On 2/26/20 1:13 PM, Bjorn Helgaas wrote:
> > On Wed, Feb 26, 2020 at 11:30:43AM -0800, Kuppuswamy Sathyanarayanan wrote:
> > > On 2/25/20 5:02 PM, Bjorn Helgaas wrote:
> > > > On Thu, Feb 13, 2020 at 10:20:15AM -0800, sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx wrote:
> > > > > From: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
> > > > > ...
> > > > > diff --git a/drivers/pci/pcie/dpc.c b/drivers/pci/pcie/dpc.c
> > > > > index 99fca8400956..acae12dbf9ff 100644
> > > > > --- a/drivers/pci/pcie/dpc.c
> > > > > +++ b/drivers/pci/pcie/dpc.c
> > > > > @@ -15,15 +15,9 @@
> > > > > #include <linux/pci.h>
> > > > > #include "portdrv.h"
> > > > > +#include "dpc.h"
> > > > > #include "../pci.h"
> > > > > -struct dpc_dev {
> > > > > - struct pci_dev *pdev;
> > > > > - u16 cap_pos;
> > > > > - bool rp_extensions;
> > > > > - u8 rp_log_size;
> > > > > -};
> > > > Adding dpc.h shouldn't be in this patch because there's no need for a
> > > > separate dpc.h file yet -- it's only included this one place. I'm not
> > > > positive a dpc.h is needed at all -- see comments on patch [4/5].
> > > Yes, if we use pdev in dpc functions used by EDR code, we should
> > > not need it.
> > I think struct dpc_dev might be more trouble than it's worth. I
> > attached a possible patch at the end that folds the 25 bits of
> > DPC-related info into the struct pci_dev and gets rid of struct
> > dpc_dev completely. I compiled it but haven't tested it.
> That would solve most of my export issues. Do you want me
> to add it to my patch list or merge it separately.
If you want to use that, I would merge it first, followed by the error
clearing cleanup, followed by the other EDR stuff.
Bjorn
